Title: Feedback, cont.
1Unit 5
- Feedback, cont.
- Goal Setting
- Schedules of Reinforcement
2Unit 5 Schedule
- Wednesday Lecture
- Monday, 10/20
- U5 lecture, cont.
- Exercise Schedules of reinforcement in the lab
and applied settings - Wednesday, 10/22 E5
- Monday, 10/27 ME1 over Units 1-4
- Ill hand out study objectives for ME1 on Monday,
10/20
3ME1
- Covers Units 1-4 (not E5)
- If you have missed an exam, you need to take ME1
or your missing exam score turns into a zero - If you have taken all the exams to date and would
like to replace your lowest score on Es 1-4, you
should take ME1 - If your grade on ME1 is lower than your grades on
Es 1-4, I throw out ME1 the ME1 cannot hurt your
grade, it can only help your grade - If you have taken all of the exams to date and
are satisfied with your scores, you get the day
off
(Monday, 11/03 is the last day to withdraw from
classes w/o academic penalty)
4In-class exercise (8 points), Mon. 10/20
- Read Latham Dossett in the coursepack
- Pay particular attention to the Methods section
- Was the continuous reinforcement (CRF) schedule
used in the Latham Dossett article a true CRF
schedule? - Was the variable ratio 4 (VR4) schedule used in
the Latham Dossett article a true VR4 schedule? - For each question, first say yes or no
- List as many reasons as you can
- You can bullet each reason, then provide an
explanation beneath each bullet
5In-class exercise (8 points), Mon. 10/22
- I am not looking for a long paper 2 pages max
- The points you earn will depend upon
- whether you identify the most obvious reasons
- the number of reasons you identify
- This is NOT an opinion paper
- Use the material from this class
- Possibly others
6SO1 SMART goals (NFE)
- Locke Latham developed goal setting theory
and have done some stellar studies in the area - While originally, Locke did not believe feedback
was important, over the years he adjusted his
opinion about that based on his research - Thus, while we would not agree with Lockes
conceptual analyses, once again, as with
expectancy theory, we do agree about practical
implementation - SMART goals (Rubin, 2002)
- Specific, Measurable, Attainable, Relevant,
Time-bound - I would add accompany with feedback and
consequences
Locke Latham (Eds.) (2013). New developments
in goal setting and task performance.
7SO2A Specific goals are better than general
goals (Locke Latham). Why from a behavioral
perspective?
- OBM position Goals affect performance only
because of the consequences that follow behaviors
that result in goal attainment. - When goals are specific
- They specify the response requirements
- They specify the criterion for reinforcement/rewar
d - Thus, both employees and managers can easily
discriminate successful from unsuccessful
performance - Goals function like task clarification in the
sense that the employee knows exactly what good
performance consists of - They also provide an explicit evaluative
component which, as I have indicated earlier,
appears to be necessary for feedback to function
effectively in most situations
(material is from an analysis by Fellner
Sulzer-Azaroff, 1984, jOBM. Time to update the
literature review, do your best goals next
evaluation component does NOT have to be goals -
could be achieved a number of ways, but goals
work )
8SO2B. What are the problems with do your best
goals?
- What about do your best goals?
- They preclude objective assessment because no
performance criteria are stated - Employees may set lower goals than the supervisor
and anticipate rewards that they then dont
receive - Remember, most employees evaluate themselves
better than their supervisor evaluates them
9SO3 Although difficult goals may lead to higher
levels of performance, be careful!
Goals should be realistic and challenging but not
too difficult From a behavioral perspective.
Why? There are 3 problems.
1. Goal R (work hard) gt Sp (supervisory
criticism)
2. Goal R (work hard) gt Ext (dont meet
goal)
3. Goal R (work hard) gt Sp (signs of
failure)
In our culture, signs of failure tend to be
conditioned punishers. Think about it.
Regardless of the criticism you get from a
supervisor (or professor), when you fail to meet
a goal or standard, how do you feel? Signs of
failure tend to be automatically punishing -
which punishes the behavior of working harder.
(example in book, students who were failing set
goals to get a 4.0 stretch goals, 2 Daniels
Oops 13 mangt practices that waste time money
operationally defined as those that are attained
less than 10 of the time students often get the
first two, miss the third 1st EOM programs)
10Jeffrey et al. study (2012) nfe
- Many organizations use one goal a goal that is
the same for all workers - Basic research question Are ability-based goals
more effective, particularly for low and middle
performers?
(lab study, interesting study, interesting
results)
11Jeffrey et al. (2012) nfe
- Based on an initial performance assessment,
divided participants into low performers, middle
performers and high performers - Two conditions
- One goal 20 of all performers had met the goal
- Ability-based goals different goals for the
different groups - Low performers 20 of low performers had met the
goal - Middle performers 20 of middle performers had
met the goal - High performers 20 of all performers had met
the goal
(based on pilot study decoding task 3-digit
number that they had to convert into a letter of
the alphabet were given the key divided Ps
into three groups based on initial session)
12Jeffrey et al. (2012) nfe
- Pay 2 base pay bonus for reaching goal
- One session, divided into 5 periods
- Practice session
- Assessment session with per piece incentives for
each item correctly decoded used to assign Ps to
low, middle, high performers - Three 5-minute performance periods
13Jeffrey et al. study (2012) nfe
- Ability-based goals were more effective than one
goal for all for low performers - Low performers who were given goals based on low
performance performed better than their
counterparts who were given a higher goal based
on the performance of all performers (which
included the middle and high performers) - Low and middle performers who received the one
goal decreased their performance across 3
sessions - Challenging, but attainable goals are the best
- What is challenging and what is too difficult??
- Were the goals in this study too difficult?
(the analysis I provided earlier probably
explains why goals were too difficult for the
low and middle performers 0 of the low and
middle performers in the one goal condition met
the goal only 35 of high performers did)
14Jeffrey et al. follow-up studies
- Goals may have been too high and bonus system may
have influenced the results - Redo the study using goals based on the average
performance of the groups, not a goal that only
20 of individuals can meet - Use piece rate pay rather than an all or none
bonus system - Urschel (2013), Replication using 3 tiered goals
versus one high goal or one medium goal with
bonus pay for reaching goals - Different perspective of traditional management
and OBM? - Only top performers should be rewarded and get
bonus pay - All workers who perform well (above average)
should be paid commensurate with their
performance (piece rate pay)
15Practical dos and donts still nfe
- When possible do use tiered (multiple goals) with
successively increasing rewards for meeting each
higher goal - Performance Matrix is a great tool for this
because of the multiple goals (columns 6-10) - Do not have different goals for different
individuals with the same tangible/monetary
rewards disaster
(organizations typically cant set individual
goals too labor intensive but it may be
possible 2-3 tiers Pampino et al. did that in
study in U2 Dans dissertation last slide on
this)
16SO5 Assigned vs participative goals
- The research indicates that participating in goal
setting does not increase performance when
compared to assigned goals - Three meta-analysis studies now support this
conclusion, first one in 1986 - The key issue, thus, is not how a goal is set,
but whether a goal is set
(Back to the Sos this is a very common
misconception issue keeps coming up at ABAI,
with a student -from another program-arguing and
maintaining that participative goals were better
just this summer I was asked about this by a
consultant who works for a behavioral firm)
17SO6A Goals what is the best combination?
- We know that goals combined with feedback are
more effective than either alone - Feedback enhances the effectiveness of goals
- Goals enhance the effectiveness of feedback
- Studies suggest that graphic feedback is the most
effective type of feedback to use with goals
better than vocal or written - We know that goals and feedback are much more
effective when consequences are provided - Not definitive, but it appears that monetary
incentives/rewards are more effective than
nonmonetary incentives/rewards - THUS.
(next slide)
18SO6A What is the best combination? (answer)
- Goals
- Graphic feedback that displays performance over
time, preferably at least once a week - Some type of performance consequence, preferably
monetary incentives
(the same recommendation from Balcazar et al
just add goals when at all possible)
19SO6B Group goals
- When using group goals, what factor should be
taken into account? - Group size
- Group goals are more effective with small groups
than large ones - However, we dont know what the critical size
is - This would be a very interesting and valuable
study to conduct - Goals, graphic feedback, consequences, while
manipulating group size -
(alone, not combining them with individual goals
group size is an issue in ALL group contingencies)
20SO7 Possible behavioral functions of goals nfe
- Analyses of goals parallel those of feedback
- Summarized the prevalent ones in SO7
- Good summary and starting point if you want to
pursue this - Tammemagi et al. (2013)
21SO8 Daniels vs. Dickinson
- Daniels maintains that if you set a goal and if
performance meets but does not exceed that goal,
the contingency is a negative rather than
positive reinforcement contingency - Also maintains that negative reinforcement
contingencies are bad contingencies because they
represent aversive control - In order for negative reinforcement to work there
must be a pre-existing aversive stimulus that the
behavior terminates or avoids - Is this a correct analysis?
22Dickinsons position
- People are not going to overshoot goals if there
is no further reinforcement for doing so, whether
the contingency is a negative or positive
reinforcement contingency - Negative reinforcement contingency
- People will perform only to the level that
terminates or avoids criticism or punishment - Positive reinforcement contingency
- People will perform only to the level that
results in maximum positive reinforcement
23Main point repeated
- If there is no further reinforcement for
performing above the goal, then people will not
exceed the goal, regardless of whether the
reinforcement is positive or negative - If you want people to perform above the goal,
then you must provide additional
reinforcement/rewards for them to do that - Daniels misconception (I think) comes from the
fact that he encourages further
praise/reward/recognition from supervisors for
above goal performance but often those rewards
are not qualitatively different than rewards for
meeting goals (nontangible socials), so doesnt
view those rewards as additional positive
reinforcement
24Example
- Union National Bank
- Baseline 1,065 items per machine hour
- Feedback 1,800 items per machine hour
- Incentive, top incentive rate was for 2,500 items
per machine hour 2,700 items per machine hour - Incentive 2, top incentive rate was for 3,500
items per machine hour 3,500 items per machine
hour - During the first incentive phase, proof operators
met but did not exceed the goal (except to a
level than insured they met the goal) - Yet during the second incentive phase when
additional incentives (reinforcement) was
provided, they increased their performance (but
again, only to a level that met the goal)
25SO9 Most common mistake re goals
- What is the most common mistake that managers
make after implementing a goal setting program
for employees? -
Increase the goals without increasing the rewards
Why is that a problem?
Its a punishment procedure. The consequence of
meeting the goal is that management increases the
goal so that the worker has to work harder and
harder to obtain the same rewards
What are employees going to do?
Restrict their productivity and in some
cases develop performance norms monitored by the
group
(mgrs loose their common sense when they become
managers. social isolation and criticism)
26What about successively increasing goals? NFE
- Daniels recommends that you set goals low so
people can meet them, then gradually increase
them - Wilk Redmon used successively increasing goals
- Sulzer-Azaroff used successively increasing goals
- Proceed with caution OK if praise and
celebrations are consequences - With tangible rewards, particularly with
incentives, you should never increase the goal
level without increasing the reward level - Tiered reward systems work well with tangible
rewards, however - Union National Bank - increased incentive rate
- Pampino et al. (U2) - an additional lottery
ticket - Performance matrix - more points for higher
levels of performance
(only after goals had been met several times
another interesting study - )
27Schedules of Reinforcement
- The basic schedules of reinforcement are
emphasized way too much in OBM. They are not very
relevant. In fact, I would argue they are not
relevant at all. Ill come back to this in a
moment - I have provided definitions of basic schedules in
SO10 (NFE) - You need to know them for your exercise this week
-
28SO11A Is hourly pay is an example of a FI
schedule?
- Yes or no
- Why or why not?
(answer not on slide)
29SO11B FR3 example, NOT!
- Goal of study increase the extent to which
college students rode buses on a college campus
(Penn State) - Tokens that could be traded for merchandise from
local stores served as the reinforcers (pop,
reduced price on pizza, etc.) - Gave a token to every third person that got on
the bus (FR3) - What is the major problem with calling this
schedule an FR3?
30Schedules of Reinforcement
- Back to Dickinsons point The basic schedules of
reinforcement tend to be emphasized way too much
in OBM. They are not very, if at all relevant. - SO12 Hantulas conclusions after reviewing the
effects of schedules of reinforcement on
organizational behavior - review covered
1971-1994 - Reinforcement schedules (in comparison to hourly
pay) are an effective way to manage work, however - The parameters of the schedule did not result in
consistent differences in performance. Rather,
the presence of a contingent relationship between
performance and rewards was the critical factor
with respect to improving performance - Bucklin Dickinson found the same thing in a
review of monetary incentives
31SO12 What does this mean?
- Performance contingent rewards do increase
performance, but different schedules of
reinforcement (e.g., FR vs VR schedules, FR1 vs
FR4, FR1 vs VR2, VR2 vs VR4) do not affect
performance differently in work settings
(ABA presentation set up incentives for staff in
human service setting - very nice study - spent
many, many hours deciding what reinforcement
schedule to use - wasted hours).
32SO13 Why are these results differentthan the
results of research on basic schedules?
- In the operant laboratory, different schedules of
reinforcement do generate different response
rates and patterns of performance. So, what may
account for the differences seen in the
laboratory and in applied settings? - Before answering, why does anyone care? Why is
this analysis important? - Our basic principles of behavior have been called
into question (particularly by expectancy
theorists in I/O) because humans do not show the
same response patterns as nonhumans - That is, they claim this proves that our basic
principles are incorrect - So, we have to be prepared to answer these
criticisms and concerns
33Two reasons why humans do not usually display the
typical performance patterns displayed by
nonhumans in an operant laboratory setting
- Although schedules used in applied settings are
indeed schedules of reinforcement, they are
rarely, if ever the same schedules examined in
the laboratory, even though they are called the
same thing (e.g., FR1, FR3, etc.). Given that
they are not the same, we should not expect the
performance patterns to be the same - FR3 example earlier
34Two reasons why humans do not usually display the
typical performance patterns displayed by
nonhumans in an operant laboratory setting
- 2. Adult humans tend to describe contingencies to
themselves and then their behavior is controlled
by their self-stated rules - FI Slow responding is reinforced
- FR Fast responding is reinforced
- Fergus Lowes (Welsh behavioral psychologist)
study with infants, 2-3 year olds and 5-year olds
(last slide on this)
35Crawley et al. article, introduction
- I have included this article because it is the
best one I have ever seen with respect to
improving sales behaviors and I would wager that
most many of the behaviors identified in their
exquisite analysis would generalize to other
sales positions - Study was conducted by one of Ed Feeneys
consultants, Bill Crawley (I didnt stress
Feeneys accomplishments in U1, but I recommend
that you go back and read the Dickinson article
for an historical perspective) - (NFE) Note the analysis at the beginning that was
designed to determine the best opportunities for
intervention, based on both the potential for
improving performance (exemplar performer vs.
average performer) and the economic pay-off of
intervening on the performance
(back to feedback old study)
36SO14 What approach was not successful in
identifying what made sales reps effective?
- Surveys were sent to the top sales
representatives in the country asking what is was
that they did that made them so effective - This approach did not work because sales
representatives could not describe the behaviors
that made them successful - I am friendly
- Its genetic - my parents were sales
representatives - You need to be up
- You need to be aggressive
- General point
- Even though workers are exemplary workers, they
often cant tell you what they do that makes them
exemplary workers. Those behaviors are often
contingency-shaped (controlled by direct-acting
contingencies) and employees never have had to
describe them (describing what you do and doing
what you do are different behavioral repertoires) - Automobile mechanics
- Construction workers
37SO15 (NFE) Exquisite specificity of the targeted
behaviors
- To determine the behaviors
- They observed 65 top performers
- Over a four month period of time
- For 1,000 hours
- Both inside the store and at in-home sales calls
- Recorded the stimulus-response sequences
- That is, what were the antecedents that prompted
a response by the sales representative, and how
did the sales representatives respond to those
antecedents - Also interviewed customers for 50 hours
- Pilot tested the entire intervention in two
stores - First with the consultant as the coach
- Then with the store manager as the coach
(fidelity - did they create an intervention that
could be carried out by employees)
38SO15 (NFE) Exquisite specificity of the targeted
areas and behaviors, cont.
- Areas with 5-11 behaviors in each area
- Customer greeted
- Customer needs identified
- Needs matched to store product and service
benefits - Objections identified and overcome
- Decision maker identified
- Close made
- Results of sales contact
- Follow-up action taken
(48 behaviors in addition to smiling, eye
contact, natural voice, and use of customers
name in each area)
39SO15 (NFE) Exquisite specificity of the targeted
areas and behaviors, cont.
- Examples of behaviors in the Customer Greeted
area - Customers should be approached within 120 seconds
after entering the store - The sales representative should stand within 3-4
feet of the customer, smile and maintain eye
contact - The sales representative should approach the
customer at a normal pace and maintain a natural
and relaxed posture - Introduce self using first and last names and
identify his/her position - Obtain the customers name and use it throughout
the sales interaction
40SO17 Two reasons why commissions did not
function as effective rewards
- Sales representatives received sales commissions
monthly which most would assume would be
sufficient to maintain high levels of performance - Commission payments were delayed, often by as
many as 3 months, weakening the relation between
sales and the amount of money earned - Commissions earned in January would not be
received until March or April - Commissions were based on sales, an
accomplishment measure, and sales representatives
did not know the behaviors required to improve
sales - The initial survey that failed to identify the
critical target behaviors showed that sales
representatives did not know what behaviors led
to improved sales
(skippingSO16 on your own Tom Gilbert, Human
comp., 1978, accomplishments v. behaviors
daniels concern, systems v. PM)
41SO18 Why is it important to compare data to
records for the same months in the preceding year?
- As part of the analysis to determine the
effectiveness of the program, they compared the
sales data to sales records for the same month
the preceding year. Why? - Sales fluctuate seasonally and monthly
- February is traditionally a big sales month while
December is traditionally a low sales month - In behavior analysis, we often use time series
data (AB design) to determine the effectiveness
of our interventions but - If you compared sales in February and it had
increased in comparison to Dec and Jan, you may
conclude that your program was successful when it
was not - Alternatively, if you compared Dec data with Nov
data (with traditionally higher sales), you may
conclude your program was not successful, when
indeed it was
(including this just so you dont just say due
to seasonal fluctuations but add an explanation)
42Crowell et al. Task Clarification
- 1. Task clarification improved performance
- 2. Objective feedback improved it further but
- 3. Praise improved it again
Task clarification alone only results in modest
increases Objective feedback should be combined
with evaluative feedback/consequences
(note that these results re objective feedback
are consistent with the Johnson article from last
unit and thus emphasizes that the difference
between objective and evaluative feedback is
important)
43NFE Crowell et al.
- My main reason for including this article was its
implications for task clarification and feedback
(see previous slide) - Secondary purpose was because of the authors
analysis of whether feedback functioned as an
antecedent, a consequence, or both (rarely done,
astute analysis) - Feedback includes task clarification, so task
clarification was examined first - If the only function of feedback was as an
antecedent, one would not expect additional
increases in performance when it was implemented
after task clarification
44NFE Crowell et al. overview
- 6 bank tellers
- 11 customer service behaviors defined
- Quality points assigned to each depending upon
importance - 100 total points possible, with 85 as the minimum
acceptable - Task Clarification
- Memo from management identifying the target
behaviors and quality points
(sos on your own)
45NFE Crowell et al. overview
- Feedback
- Posted chart with individual point scores, daily,
coded - Fake scores for any teller that was absent to
protect confidentiality - Supervisor met with each teller individually when
the teller came to work, but descriptive not
evaluative - Praise added
- Supervisors praised tellers if the point score
was above 85 or below 85, but higher than the
preceding score
46NFE Crowell et al. results
Phase Average
Baseline 1 61.4 points
Task clarification 1 72.0 points
Feedback 1 78.0 points
Feedback praise 1 83.0 points
Baseline 2 76.0 points
Feedback 2 83.0 points
Feedback praise 2 88.0 points
last 6 sessions, gt 85
only phase mean gt 85
Above standard performance was obtained and
sustained only when both feedback and praise were
added to task clarification
47NFE Social validity not definitive, but actual
hard datafacts, not survey data cool!
- Dollars on deposit in bank
- 24 to 42 million, 75 growth
- Customer complaints
- 2-3 per month, dropped to near-zero level
- Compliments increased
(no customer input, so no certainty customer
perceptions and behaviors were affected by the
intervention, but results continued on next
slide)
48NFE Social validity not definitive, but actual
hard datafacts, not survey data cool!
- Bank managers rolled-out the program
- Streamlined and implemented it in 6 other
branches - Program was being implemented in the remaining 5
branches at the time the study ended - Cost
- Annual cost of program as implemented 6,000.00
- Streamlined version in other 11 branches, an
additional 16,000 - This included the cost of a dedicated full-time
program administrator/observer
(streamlining next slide)
49NFE Social validity not definitive, but actual
hard datafacts, not survey data cool!
- Streamlined version
- Reduced teller observations from daily to 3X a
week - Reduced the number of transactions recorded per
teller in each session from six to three - Once praise was introduced, a maintenance
procedure was suggested in which the frequency of
recording be further reduced to one session per
week
(reduce labor intensiveness and number of
observers required ok last slide on this, moving
on)
50Gaetani Johnson, cash shortages, intro
- Purpose
- Decrease cash and inventory shortages in a
retail beverage chain (employee theft?) - Assessed the effects of data plotting
(self-recording), praise, and state lottery
tickets - Im including it because of the comparison of
self-recording/praise/lottery tickets and as
another example of a lottery system not
expensive but effective - Unusual lottery system in that each mgr who met
the criterion received 2 state lottery tickets
(only 1.00 a piece!) - Most lottery systems employees who meet the
criterion have a chance to win a prize, but same
basic principle -
(also including it tie in with the Johnson
study from last unit address that in a moment)
51Gaetani Johnson, cash shortages, intro
- Participants 12 store managers
- DV Efficiency estimate
- Very nice measure could be computed from data
the stores were already collecting and equated
high sales volume and low sales volume stores - Researchers did not implement any new measurement
system or data collection system BIG PLUS - Calculate baseline measures using archival data
- IVs
- Data plotting alone (closed circles on graph)
- Praise alone (open triangles on graph)
- Data plotting and praise (closed triangles on
graph) - Data plotting, praise, and lottery tickets
(closed squares on graph)
(one of few studies that has examined supervisory
praise alone without specific graphic feedback
weekly phone conversations this is the
tie-in with the Johnson study from last week
praise alone without objective feedback)
52SO29 Least effective and most effective IVs
- Least effective
- Data plotting alone
- Supervisory praise alone
- Most effective
- Data plotting, praise, and lottery tickets
- Group A stores, 123 above baseline
- Group B stores, 108 above baseline
- Group C stores, 40 above baseline
- Group C, no prior exposure to self-recording or
praise as Group A and Group B managers, so prior
exposure may well have affected the results
(skipping to SO29, are straightforward careful
with wording data plotting praise better)
53NFE, but remember your consequences!
- Self-monitoring alone does not seem to be a
particularly powerful intervention - Richman et al. study, look at in U8
- Gaetani et al. (1983) The study I recommended
but did not include in coursepack - Supervisory praise by itself does not seem to be
a powerful intervention (need objective feedback
as well) and objective feedback itself may not
be powerful w/o praise or some type of evaluation
(Crowell study) - Once again, the combination of feedback, praise,
and tangible rewards appears to be the most
powerful
(note the consistency of these results with
Johnson- thats why I like that article so much
it ties in and Explains the results of some of
these applied studies more controlled study)
54Wilk Redmon article
- Study was conducted as Braksicks (Wilks)
doctoral dissertation while she was at WMU - Excellent model of how to do research in the real
world few better examples - Follow-up of a study conducted at WMU in our
admissions and orientation office - Pam Liberacki, Director of Admissions and
Orientation, who retired three years ago - Braksick was hired as a consultant to implement
the program at UM based on the success of the
program here
(not going to go over many of the Sos one of
CLGs founders in 1993 both she and Julie Smith
have now left this study was Excellent predictor
of her future work excellent, excellent work)
55Wilk Redmon intro, cont.
- Participants were 16 clerical workers at UM
- DVs
- Number of tasks completed
- Performance efficiency
- Employee satisfaction
56SO32 Why was the efficiency measure used?
- Performance efficiency formula (NFE)
- Total number of tasks completed by all
participants - Total number of hours worked by all
participants - Why is this an important measure - why not just
use the total number of tasks completed? - The total number of hours worked by the employees
differed from week to week - If you only looked at the total number of tasks
completed, you wouldnt know whether workers were
completing more tasks because they were working
more hours or whether they were completing the
more tasks in the same amount of time - If workers completed more tasks but also worked
more hours, then you have not increased
performance
57Skipping to SO37 What procedure was used to
verify that the supervisor actually delivered the
feedback?
- After feedback was given during the week day, the
employee placed a check mark on the next entry on
their data sheet - If you use a graphic feedback display, have
employees initial the graphic feedback display - If you post a graph, have employees initial the
posted graph - More modern technology send the graph or
feedback via email with verification that the
email has been opened by the recipient - not as good - employees could conceivably open
the email and not look at the feedback, but
better than nothing
(I am pointing this out because it is an
excellent procedure - its simple, doesnt
require any extra effort on the part of the
researcher, yet does confirm that feedback was
provided as it was supposed to be provided -
fidelity of implementation of the IV - a lot of
our students at WMU have used this or something
similar in their studies )
58(Results! circled data are an issue decreasing
trend back to previous level but others, large,
abrupt change average data next)
59SO40 Most importantly, what does this study
reveal?
Line
The important role that graphic feedback plays in
improving performance
(average data click highlight last slide on
Wilk except questions, comments)
60Questions over Wilk and Redmon?
61NFE Take home points about goals and feedback
- Ability-based/individualized goals are the most
effective (Jeffrey et al., 2012) - Probably helps account for dynamite results in
Wilk Redmon, 1993 - Objective plus evaluative feedback (praise)
better then either alone (Johnson, 2013 Crowell
et al., 1988 Gaetani Johnson, 1983) - Graphic feedback is better than vocal feedback
(Wilk Redmon, 1993)
62In-class exercise
- Was the continuous reinforcement (CRF) schedule
used in the Latham Dossett article a true CRF
schedule? - Was the variable ratio 4 (VR4) schedule used in
the Latham Dossett article a true VR4 schedule?